Innovator: Learn from Fermi!

By Shlomo Maital


Enrico Fermi 1901-1954

   One of the key problems innovators encounter is validation. You have an idea. You love it! You’re full of passion to implement it. But, hold it. Take a moment or more, and make sure you’re meeting a real unmet need. Validate the need. Don’t be like so many entrepreneurs who “fill a much-needed gap”.

   But how? How to validate? Talk to people? Sure. But will chatting with 1, 2 or even 10 people do the trick?

   I’ve just learned about an interesting method that may help. (See: Philip Tetlock, Dan Gardner, Superforecasting: Crown, 2015; see Fermi-ize Your Problem). It comes from the Italian physicist Enrico Fermi, who died young at 53; he was an exceptional scholar, both a theoretical and experimental physicist (very rare to do both), and invented the first nuclear reactor, Chicago-Pile 1, in Chicago. He won the Nobel Prize in 1938, and has an element named after him. (Fermium, element 100, symbol Fm).

   Here is how Fermi taught his students to tackle tough problems. I’ve slightly changed the context.

   Suppose you want to create an Android app for tuning pianos. Suppose you want to sell it in Chicago. How can you validate the need? How can you predict how many you will sell?

   Size of market: How many piano tuners are there in Chicago?

   No possible way I can find that number on the Internet. Impossible task. Dead end.

     Not according to Fermi.   Take a hard question. Break it down into pieces. And this will enable you to come up with a very good guess that will be surprisingly accurate.

For instance: 1. How many pianos are there in Chicago?   Well, the population of Chicago is 2.5 million. Let’s say, one person in 100 has a piano. (about 1 household in 25 or 30). That means: 25,000 pianos. But schools, etc. also have pianos. That could be, say, another 25,000 pianos. So there are 50,000 pianos. 2. How often are pianos tuned? Let’s say, once a year. Reasonble, right? 3. How long does it take to tune a piano? Roughly, two hours. Remember: each note has to be played, tested, checked. And a piano has 88 keys, 52 white and 36 black. 4. How many hours per year does a piano tuner work? Let’s say, 40 hours a week, for 50 weeks, or 2,000 hours. But take off 20 per cent for travel time: so, 1,600 hours.

     Final calculation: Total piano tuning time: 50,000 pianos times 2 hours =   100,000 hours per year. Divide by 1,600 hours per tuner:   You get about 60 or so piano tuners.

     Does that comprise a sufficient market for your app? Chicago alone, no. So you need to be nationwide or global. Can you be? How will they know about your app?

     This is called Fermi-ize Your Problem. Take a really hard problem. Break it down into pieces. Tackle each piece.

     What if your assumptions are wrong? Well – if you’ve written them down, you can check them. So – instead of giving up from the outset when tackling a hard question, you break it into pieces and tackle each piece,   and when you solve each piece, you chalk up progress and gain momentum. And move one step closer to an answer.

     Innovator: Try to Fermi-ize your problem. Fermi did this regularly with his students. I think it helped him win a Nobel Prize. If you get good at Fermi-izing, you become more willing to tackle really tough problems..and change the world. And you may avoid investing years solving a problem no one really cares about.